America's Next 'Stop Model!': Model Disgorgement

Abstract

This Article explores the emergence of model disgorgement–the compelled destruction or dispossession of certain data, algorithms, models, and associated work products created or shaped by illegal means--as a remedy, right, and requirement for artificial intelligence and machine learning systems. Part I examines model disgorgements emergence as a consumer protection remedy and conception as a positive right and regulatory requirement. Part II considers the constellation of Federal and State actors who might seek model disgorgement to address particular AI and ML harms and violations of law, including Federal and State enforcement agencies and legislative bodies. Part III underscores the need to legislate broader privacy and data protection regulation to bolster the effectiveness of model disgorgement. Part IV reflects on the challenges of model disgorgement, including the scope and enforcement of model disgorgement orders and logistical issues for companies undergoing model disgorgement. Part V envisions two speculative case studies for model disgorgement’s application: Clearview and Securus.

Keywords: artificial intelligence, machine learning, consumer protection, disgorgement, FTC, model